Remaining Useful Life Estimation Research Articles

A novel dual-stream self-attention neural network for remaining useful life estimation of mechanical systems

Remaining useful life (RUL) estimation plays a crucial role in evaluating health states and improving maintenance plans of mechanical systems. Recently, artificial intelligence-based data-driven methods that use monitoring data as input have made significant progress in machine prognostics. However, current methods commonly ignore the correlations and internal differences of monitoring data, consequently leading to limited estimation performance. Therefore, this paper proposes a novel data-driven RUL estimation method named Dual-Stream Self-Attention Neural Network (DS-SANN). First, the multi-head self-attention mechanism is employed to learn correlations between different monitoring data and weigh the features dynamically to obtain global degraded information. Then, a dual-stream structure network is established to extract features from the original and auxiliary data simultaneously to make a comprehensive reflection of health states. The original and auxiliary data represent absolute values and internal differences of monitoring data, respectively. Finally, the multilayer perceptron is adopted to fuse the obtained features and estimate RUL. In addition, the effectiveness of DS-SANN is validated by the public degradation dataset of turbine engines. Compared with several existing prognostics methods, DS-SANN shows better estimation performance when averaging across all sub-datasets. Specifically, estimation effects evaluated by RMSE and Score improve 21.77% and 32.67%, respectively.

On-board vibration-based remaining useful life estimation of hollow worn railway vehicle wheels

Abstract The feasibility of on-board remaining useful life (RUL) estimation of hollow worn railway vehicle wheels under varying travelling speed using vibration signals from the vehicle’s bogie, is for the first time explored. A variant of the Multiple Model Power Spectral Density (MM-PSD) method is employed, treating RUL estimation as a multi-class classification problem, where each class corresponds to a specific condition of hollow worn wheels associated with an empirical remaining mileage. The method classifies the current, unknown, condition of hollow worn wheels to one of the available by examining the vehicular dynamics similarity among them through the amplitude of the sharp periodic valleys (antiresonances) induced due to the Wheelbase Filtering (WF) effect. The vibration signals are obtained from Monte Carlo experiments based on the high-fidelity SIMPACK software, while the remarkable performance of the method is demonstrated via thousands of inspection cases.

DOWELL: Diversity-Induced Optimally Weighted Ensemble Learner for Predictive Maintenance of Industrial Internet of Things Devices

The Industrial Internet of Things (I-IoT) enables a smarter maintenance approach for various industrial applications, such as manufacturing, logistics, etc. This approach is based on continuously observing system data to predict device failures and increase device efficiency. This smart maintenance, also known as predictive maintenance (PDM), finds an optimal maintenance schedule to reduce operational and capital costs. Accurate remaining useful life (RUL) prediction is critical for an effective PDM system. Data-driven RUL estimation methods are quite popular owing to their easier implementation. We observe that the performance of data-driven methods varies drastically based on the data set and underlying system parameters, thus making it difficult to have a single algorithm and a parameter set that work best for all settings. We propose an ensemble learning framework, where accurate and diverse base learners are selected out of 20 different state-of-the-art deep learning models. For accuracy, we discover the optimal weights of base learners by constructing an optimization problem. For diversity, we measure the similarity among base learner predictions and iteratively select the most diversified set of models while keeping the accuracy at a certain level. We show that our approach can have 39.2% faster retraining compared to an accuracy-based ensemble with only 3.4% loss in accuracy.

Aircraft engine remaining useful life estimation via a double attention-based data-driven architecture

Remaining useful life (RUL) estimation has been intensively studied, given its important role in prognostics and health management (PHM) of industry. Recently, data-driven structures such as convolutional neural networks (CNNs), have achieved outstanding RUL prediction performance. However, conventional CNNs do not include an adequate mechanism for adaptively weighing input features. In this paper, we propose a double attention-based data-driven framework for aircraft engine RUL prognostics. Specifically, a channel attention-based CNN was utilized to apply greater weights to more significant features. Next, a Transformer was used to focus attention on these features at critical time steps. We validated the effectiveness of the proposed framework on benchmark datasets for aircraft engine RUL estimation. The experimental results indicate that the proposed double attention-based architecture outperformed the existing state-of-the-art (SOTA) algorithms. The double attention-based RUL prediction method can detect the risk of equipment failure and reduce loss.

Cross-condition and cross-platform remaining useful life estimation via adversarial-based domain adaptation

Supervised machine learning is a traditionally remaining useful life (RUL) estimation tool, which requires a lot of prior knowledge. For the situation lacking labeled data, supervised methods are invalid for the issue of domain shift in data distribution. In this paper, a adversarial-based domain adaptation (ADA) architecture with convolution neural networks (CNN) for RUL estimation of bearings under different conditions and platforms, referred to as ADACNN, is proposed. Specifically, ADACNN is trained in source labeled data and fine-tunes to similar target unlabeled data via an adversarial training and parameters shared mechanism. Besides a feature extractor and source domain regressive predictor, ADACNN also includes a domain classifier that tries to guide feature extractor find some domain-invariant features, which differents with traditional methods and belongs to a unsupervised learning in target domain, which has potential application value and far-reaching significance in academia. In addition, according to different first predictive time (FPT) detection mechanisms, we also explores the impact of different FPT detection mechanisms on RUL estimation performance. Finally, according to extensive experiments, the results of RUL estimation of bearing in cross-condition and cross-platform prove that ADACNN architecture has satisfactory generalization performance and great practical value in industry.

A new methodology for estimation of dynamic Remaining Useful Life: A case study of conveyor chains in the automotive industry

In the context of industry 4.0 and digital transformation, the predictive maintenance (PdM) approach plays an important role in the efficiency of maintenance, as breakdowns and over-maintenance. Remaining Useful Life (RUL) is the main part of the prognostic aspect of maintenance. RUL is also one key piece of information that feeds the PdM and needs to be dynamically readjusted in a global approach and procedure. In this paper, a new methodology for the prognostic aspect of PdM, and a dynamic RUL estimation method have been proposed. In this way, general methodologies and processes of prognostics have been presented that outline a more coherent vision toward RUL estimation. Within the dynamic RUL method, it is proposed to use from Prophet prediction model in a dynamic algorithm to better estimation of RUL based on the Health Indicator (HI) trends updates. The applicability and efficiency of the proposed procedure and method have been applied and validated in the conveyor chains of an automotive company which their failure results in production stoppages and significant damages. The performance of the prediction method has been presented with a comparison to deep learning and statistical prediction methods. Following the dynamic RUL estimation, a maintenance strategy has been proposed for the studied case to improve maintenance planning.

Dual-Attention-Based Multiscale Convolutional Neural Network With Stage Division for Remaining Useful Life Prediction of Rolling Bearings

Remaining useful life (RUL) prediction of rolling bearings is of great importance in improving the reliability and durability of rotating machinery. This paper proposes a dual-attention-based convolutional neural network with accurate stage division for rolling bearings RUL prediction, which includes two subsections, i.e., First prediction time (FPT) determination and RUL estimation. Firstly, signal features characterizing the bearing degradation process are fused by Wasserstein Distance to perform two stage division with great robustness. The correct labeled RUL samples with explicit degradation property are then prepared for future network training. Dual attention mechanism is adopted to not only focus on the effect of different sensor signals but also different time steps. Afterwards, multiscale convolution is utilized to both extract local and global weighted features to obtain more comprehensive information. Finally, several convolutional blocks are applied to further obtain accurate RUL prediction. The results derived from fault-mechanism-based simulation signals and experimental signals show that the proposed method is more effective and robust by ablation and comparison study.

Aircraft Engines Remaining Useful Life Prediction Based on A Hybrid Model of Autoencoder and Deep Belief Network

Remaining Useful Life (RUL) is used to provide an early indication of failures that required performing maintenance and/or replacement of the system in advance. Accurate RUL prediction offers cost-effective operation for decision-makers in the industry. The availability of data using intelligence sensors leverages the power of data-driven methods for RUL estimation. Deep Learning is one example of a data-driven method that has a lot of applications in the industry. One of these applications is the RUL prediction where DL algorithms achieved good results. This paper presents an Autoencoder-based Deep Belief Network (AE-DBN) model for Aircraft engines’ RUL estimation. The AE-DBN DL model is utilized the feature extraction characteristic of AE and superiority in learning long-range dependencies of DBN. The efficiency of the proposed DL algorithm is evaluated by comparison between the proposed AE-DBRN and the state-of-the-art related method for RUL perdition for four datasets. Based on the Root Mean Square Error (RMSE) and Score indices, the outcomes reveal that the AE-DBN RUL prediction model is superior to other DL approaches.

Attention-based CNN-BiLSTM for SOH and RUL estimation of lithium-ion batteries

The state of health (SOH) and remaining useful life (RUL) of lithium-ion batteries describe the current aging degree of the batteries from different perspectives, and accurate and efficient battery health estimation is essential for their safe use. To improve the effectiveness and accuracy of the batteries’ health assessment models, this paper proposes a new method for SOH and RUL estimation of lithium-ion batteries. Convolutional neural networks (CNNs), bi-directional long short-term memory (BiLSTM), and attention mechanism (AM) to build a hybrid network model for capacity estimation of lithium-ion batteries, and further calculate the SOH and RUL estimation results. By using Center for Advanced Life Cycle Engineering (CALCE) lithium-ion battery capacity degradation data, we extracted the battery health indicator (HI) and verified the reasonableness of HI selection by using Gray Relational Analysis (GRA) and compared it with other network models to calculate the prediction accuracy by various evaluation indexes. The experimental results show that the method has higher estimation accuracy while avoiding the construction of complex battery mechanism degradation models and is highly generalized.

Convolutional Transformer: An Enhanced Attention Mechanism Architecture for Remaining Useful Life Estimation of Bearings

Nowadays, deep learning (DL) methods for prognostic and health management (PHM) have vastly broadened the scope of applications in this field. Numerous approaches based on deep neural networks have been presented and applied to the remaining useful life (RUL) estimation of bearings. However, few of these methods are yet fully competent for the task of extracting degradation-related information from raw signals both locally and globally. To fill this research gap, we proposed a novel convolutional Transformer (CoT) which combines the global context capturing of attention mechanism with the local dependencies modeling of convolutional operation. Specifically, we designed a multi-scale convolutional (MSC) module with Swish activation for Transformer architecture to embed local feature learning into global sequence modeling. Our CoT fuses intra-token convolution and inter-token self-attention operations to enable simultaneous extraction of local dependencies and global interactions from the raw temporal signal into a trainable class token. Then, an end-to-end RUL estimation framework based on CoT is presented, which provides a mapping from raw vibration signals to estimated RULs. Finally, comprehensive case studies, including comparative studies and ablation experiments, fully validate the effectiveness and advancements of our CoT-based RUL estimation method.

Remaining Useful Life Estimation for Ball Bearings Using Feature Engineering and Extreme Learning Machine

Rotating machines, such as pumps and compressors, are critical components in refinery and chemical plants used to transport fluids between processing units. Bearings are often the critical parts of rotating machinery, and their failure could result in economic loss and/or safety issues. Therefore, estimation of the remaining useful life (RUL) of a bearing plays an important role in reducing production losses and avoiding machine damage. Because bearing failure mechanisms tend to be complex and stochastic, data-driven RUL estimation approaches have found more applications. This work proposes a novel RUL estimation method based on systematic feature engineering and extreme learning machine (ELM). The PRONOSTIA dataset is used to demonstrate the effectiveness of the proposed method.

A Hybrid Model-Based Approach on Prognostics for Railway HVAC

Prognostics and health management (PHM) of systems usually depends on appropriate prior knowledge and sufficient condition monitoring (CM) data on critical components’ degradation process to appropriately estimate the remaining useful life (RUL). A failure of complex or critical systems such as heating, ventilation, and air conditioning (HVAC) systems installed in a passenger train carriage may adversely affect people or the environment. Critical systems must meet restrictive regulations and standards, and this usually results in an early replacement of components. Therefore, the CM datasets lack data on advanced stages of degradation, and this has a significant impact on developing robust diagnostics and prognostics processes. This paper proposes a methodology for implementing a hybrid model-based approach (HyMA) to overcome the limited representativeness of the training dataset for developing a prognostic model. The proposed methodology is evaluated building an HyMA which fuses information from a physics-based model with a deep learning algorithm to implement a prognostics process for a complex and critical system. The physics-based model of the HVAC system is used to generate run-to-failure data. This model is built and validated using information and data on the real asset; the failures are modelled according to expert knowledge and an experimental test to evaluate the behaviour of the HVAC system while working, with the air filter at different levels of degradation. In addition to using the sensors located in the real system, we model virtual sensors to observe parameters related to system components’ health. The run-to-failure datasets generated are normalized and directly used as inputs to a deep convolutional neural network (CNN) for RUL estimation. The effectiveness of the proposed methodology and approach is evaluated on datasets containing the air filter’s run-to-failure data. The experimental results show remarkable accuracy in the RUL estimation, thereby suggesting the proposed HyMA and methodology offer a promising approach for PHM.

Remaining Useful Life Prediction of Machinery: A New Multiscale Temporal Convolutional Network Framework

With the rapid development of deep learning (DL) techniques, data-driven models have been increasingly used in remaining useful life (RUL) prediction, in which convolution neural network (CNN)-based RUL prognostics models have received special attention. However, there are still two main issues that need to be addressed: (1) Traditional CNN is not suitable to extract the time-sequence characteristics from the long-term historical signals. (2) The receptive field range of convolution operation is fixed, thus only learning the feature information at a specific scale, which is insufficient for complex feature extraction. To address these two issues, a multi-scale temporal convolutional network (MsTCN) which has powerful time-sequence characteristics is proposed for RUL prediction in this paper. The MsTCN adopts the temporal convolutional network (TCN) framework, which is good at extracting time-sequence information. Based on this, a new multi-scale dilated causal convolution residual block (MsDCCRB) is developed to constitute the RUL prognostics model, where multiple dilated convolutions (DCs) are based on different dilation factors are put on each layer in parallel. Furthermore, the squeeze-and-excitation (SE) unit is embedded into the MsDCCRB to adaptively recalibrate the sequence feature responses and enhance the representation learning ability. Through stacking multiple MsDCCRBs, the historical condition monitoring data can be fed directly into the proposed model to realize the high-level representations of RUL estimation. Finally, the proposed approach is validated with the accelerated whole-life degradation dataset of rolling element bearings (REBs). The experimental results exhibit the proposed MsTCN achieves a higher RUL prediction accuracy, which is superior to some state-of-the-art data-driven prognostics methods.

A Remaining Useful Life Model for Optimizing Maintenance cost and Spare-parts replacement of Production Systems in the Context of Sustainability

In reliability and maintenance engineering, predictive methods play a major role for estimating the remaining useful life (RUL) of equipment. However, in the most cases the RUL estimation is based on two types of approaches: model-based solution and data-driven solution. Data-driven solution is a very realistic solution, but requires the availability of large quantity of collected data using sensors networks, big storage capacity, supercomputers for processing, and high-level algorithms such as machines learning, convolutional Neural Networks, Hidden Markov Models. While model-based solution methods are less difficult to deploy, they use techniques based on the mean residual life (MRL) value. This paper proposes an integrated decision support model for minimizing the maintenance expected cost, and for reducing the excess of spare-parts usage of multi-component production systems. This decision support model includes three performance indicators that are: the renewal function (RF), the mean residual lifetime (MRL) and the renewal mean residual lifetime (RMLR). The contribution of this research consists of (1) introducing the MRL function for a class of useful probability distributions, (2) proposing the Renewal MRL (RMRL) as a predictive maintenance strategy, (3) applying the approach-MRL to maintenance scheduling problems, (4) dressing a comparative numerical study between the different proposed models using an industrial case study. Besides, the comparative study performing the RF, MRL, and RMRL is carried out using basic and modified replacement strategies consisting of an age replacement policy (ARP) and an opportunistic maintenance solution. In addition, a case study from an Electrical power system is proposed by providing numerical results that discuss and illustrate the outcome gains in terms of maintenance costs saving and spare parts replacements’ minimization. However, the proposed solution is demonstrated to have a positive effect to enhance sustainable development systems, and it is provided soon to cover other theoretical and application aspects of the MRL.

A Deep Learning Model for Remaining Useful Life Prediction of Aircraft Turbofan Engine on C-MAPSS Dataset

In the era of industry 4.0, safety, efficiency and reliability of industrial machinery is an elementary concern in trade sectors. The accurate remaining useful life (RUL) prediction of an equipment in due time allows us to effectively plan the maintenance operation and mitigate the downtime to raise the revenue of business. In the past decade, data driven based RUL prognostic methods had gained a lot of interest among the researchers. There exist various deep learning-based techniques which have been used for accurate RUL estimation. One of the widely used technique in this regard is the long short-term memory (LSTM) networks. To further improve the prediction accuracy of LSTM networks, this paper proposes a model in which effective pre-processing steps are combined with LSTM network. C-MAPSS turbofan engine degradation dataset released by NASA is used to validate the performance of the proposed model. One important factor in RUL predictions is to determine the starting point of the engine degradation. This work proposes an improved piecewise linear degradation model to determine the starting point of deterioration and assign the RUL target labels. The sensors data is pre-processed using the correlation analysis to choose only those sensors measurement which have a monotonous behavior with RUL, which is then filtered through a moving median filter. The updated RUL labels from the degradation model together with the pre-processed data are used to train a deep LSTM network. The deep neural network when combined with dimensionality reduction and piece-wise linear RUL function algorithms achieves improved performance on aircraft turbofan engine sensor dataset. We have tested our proposed model on all four sub-datasets in C-MAPSS and the results are then compared with the existing methods which utilizes the same dataset in their experimental work. It is concluded that our model yields improvement in RUL prediction and attains minimum root mean squared error and score function values.

Aero-Engine Remaining Useful Life Estimation Based on Multi-Head Networks

Data-driven aero-engine remaining useful life (RUL) estimation is a key technology to monitor engine’s degradation. However, due to the difficulties of extracting the time information from data, the accuracy of data-driven methods remains low. Aiming at the problem, this article proposes a novel multi-head structure to wrap time information into the network structure and training processes. Multi-head networks are designed to take the time sequence information of inputs in the feedforward process. Meanwhile, the time loss function takes full consideration of the time prior information of the outputs (estimated RUL) and directs the backward process of the networks to converge to the real aero-engine operating situation. Experiments on the NASA commercial modular aero-propulsion system simulation (C-MAPSS) aero-engine’s RUL estimation dataset are conducted to validate the effectiveness of the proposed method. The result and comparisons with other state-of-the-art methods show that the proposed multi-head structure and time loss function can improve the accuracy significantly. This suggests that the proposed method is a promising approach in aero-engine degradation evaluation.

Empirical Analysis for Remaining Useful Life Estimation via Data-Driven Models

Today, due to the growing complexity of engineered systems, it is crucial to develop technologies able to deal with the systems' behavior to maintain a high degree of safety, reliability, and efficiency while reducing operating expenses such as maintenance costs. The idea is to develop a Prognostics and Health Management (PHM) framework to monitor these complex systems' behavior using sensory data and then apply machine learning models to infer the current health state. One important goal is to estimate the Remaining Useful Life (RUL) essential for optimizing maintenance processes and sustainable practices in industrial settings. This work presents an empirical analysis for RUL estimation via a model degradation built using condition monitoring data. A support vector machine regression (SVR) model and a similarity measure algorithm are employed to extract degradation trends and compute the RUL. We evaluate the prognostics performance and compare the results with reported benchmarks from publicized works.

A Nonlinear Prognostic Model Based on the Wiener Process with Three Sources of Uncertainty

Estimation of the remaining useful life (RUL) is an important component of prognostics and health management (PHM). The accuracy of the RUL estimation for complex systems is mainly affected by three sources of uncertainty, i.e., the temporal uncertainty, the product-to-product uncertainty, and measurement errors. To improve PHM and account for the effects of the three sources of uncertainty, a nonlinear prognostic model with three sources of uncertainty is presented here. An approximated analytical expression for the probability density function (PDF) of the RUL is obtained based on the concept of first hitting time (FHT). Model parameters are then obtained by the expectation maximization (EM) algorithm, and the drift parameter is estimated adaptively using a Bayesian procedure. Finally, in order to illustrate the practical applications of the presented approach, a comparative study of real data on fatigue crack propagation is presented. Results demonstrate that our method improves model fit and increases the accuracy of the lifetime estimation.

An improved prognostics model with its application to the remaining useful life of turbofan engine

Machinery prognostics play a crucial role in upgrading machinery service and optimizing machinery operation and maintenance schedule by forecasting the remaining useful life (RUL) of the monitored equipment, which has become more and more popular in recent years. The safety of aviation is one of the issues that people are most concerned about in the field of transportation, since it might cause disastrous loss of life and property once accident happened. The turbofan engine is an important part of the aircraft that provides thrust for plane. With aging, the turbofan engine becomes prone to failures. As a result, it would be worth studying prognostics in turbofan engine to improve the reliability of machinery and reduce unnecessary maintenance cost. Recently, a data-driven prognostics modeling strategy called the classification of predictions strategy (CPS) was proposed, in which the continuous signal and the discrete modes of an actual system come together to achieve RUL estimation. However, machine health states measured from classification rarely have just one potential situation, and this strategy cannot determine whether the fault occurs or not by a certain probability which comes closer to reality. Moreover, since there is no information and prior knowledge of prognostics application, it is hard to obtain the probability of various situations from raw measured data. Hence, based on previous work, this paper proposes an improved prognostics modeling method named the classification of predictions strategy with decision probability (CPS-DP), whose key innovations mainly include three parts: (1) decision probability process (DPP) where each step of multi-step prediction obeys geometric distribution and can judge whether the failure state occurs using the decision probability; (2) decision probability calculation (DPC) algorithm, which is first proposed by this paper and can calculate the values of decision probability without prior knowledge of prognostics application; and (3) withdrawal mechanism optimizer (WMO), which is specially designed to compensate for the shortcomings of DPP and further enhance the performance of the prognostics model. In brief, first, CPS is used to build a basic prognostics model to acquire RUL estimation results, in which the information applied to find the probability has been contained. Later, the mean of RUL estimation errors is figured from the results, which is further employed to calculate the probability using DPC algorithm. Then, CPS-DP can be achieved by means of integrating two parts: DPP and CPS. Furthermore, to further improve the performance, WMO is utilized to optimize CPS-DP with rolling back predictions. Ultimately, an enhanced prognostic model based on CPS-DP is set up through uniting CPS, DPP, and WMO. To validate the proposed method, experimental results on the turbofan engine in 2008 prognostics and health management competition are investigated.

Bayesian Optimization LSTM/bi-LSTM Network With Self-Optimized Structure and Hyperparameters for Remaining Useful Life Estimation of Lathe Spindle Unit

Abstract An effective maintenance strategy to cut back maintenance costs and production loss with assured product quality has always been a major concern for industries. The Industry 4.0 era has built a wide acceptance for the predictive maintenance techniques in the remaining useful life (RUL) estimation of critical industrial systems. In this paper, long short-term memory (LSTM) and bidirectional-LSTM (bi-LSTM) deep neural architecture-based predictive algorithms are proposed for the RUL estimation of the lathe spindle unit. The deep learning algorithm is embedded within a Bayesian optimization algorithm for the self-optimization of its network structure and hyperparameters. The proposed deep learning algorithm is trained using lathe spindle health degradation data collected from an experimental accelerated run-to-failure test rig to evolve an RUL prediction model. The vibration signals representing lathe spindle health degradation from the health to faulty state are analyzed to extract time, frequency, and time-frequency domain features, which are then subjected to a neighborhood component analysis (NCA) based feature selection criteria. Finally, the selected relevant features are used to train the optimized LSTM/bi-LSTM network for RUL estimation. A comparison of the prediction results for Bayesian optimized LSTM/bi-LSTM network architectures and other prominent data-driven approaches are performed. The Bayesian optimized LSTM + bi-LSTM deep network architecture is observed to have the highest prediction accuracy for lathe spindle RUL estimation.